ESPN: "Real" Plus-Minus

[Wannabe MEP]

I've been looking at variations of this stat for a few years now…I don't think this one is a dramatic new revelation. But now it has ESPN cred.

http://espn.go.com/nba/story/_/id/10740 ... plus-minus
http://espn.go.com/nba/statistics/rpm/_/sort/RPM

I do think it is the best single-number measure out there, but there is no perfect single-number measure. It's attempting to strip out context, but that's impossible. Fit matters. So this needs to be analyzed alongside lineup data, other advanced stats, and of course the eye-test (and pure basketball knowledge).

Here are the Suns' ratings:

20) Frye +4.06
26) Dragic +3.69
31) Bledsoe +3.52
120) Shavlik +0.83
151) Marcus +0.06
154) Markieff +0.01
182) Tucker -0.53
193) Green -0.77
231) Barbosa -1.41
239) Ish -1.51
253) Plumlee -1.66
328) Christmas -2.64
356) Len -3.15

A few thoughts on this…
1) 0 is not a bad score -- it's meant to be "average" for an NBA player. Slight negative isn't even bad, considering role and price. Amare at -5.84 and $21.7 million? THAT'S bad.
2) Low minute players are always hard to figure out, so I wouldn't take this to mean too much for those guys.
3) I don't know how they're doing priors on this. Engelmann has used slight negative values as the priors for rookies, so that may be the case here. I also don't know how many years of data they're using, although it appears to be weighted heavily for this season.

[Ring_Wanted]

There is a substantial thread about this start on the general board, with good explanations. Some apparent aberrations like Nick Collison and Channing Frye have been discussed too, so you might want to check it out.

[Wannabe MEP]

There is a substantial thread about this start on the general board, with good explanations. Some apparent aberrations like Nick Collison and Channing Frye have been discussed too, so you might want to check it out.

Thanks for the tip. Take a look, people: viewtopic.php?f=6&t=1314113

Definitely take a look at Doctor MJ's post on low-minutes guys like Collison on page 5: viewtopic.php?f=6&t=1314113&start=80

[RunDogGun]

There is a substantial thread about this start on the general board, with good explanations. Some apparent aberrations like Nick Collison and Channing Frye have been discussed too, so you might want to check it out.

Thanks for the tip. Take a look, people: viewtopic.php?f=6&t=1314113

Definitely take a look at Doctor MJ's post on low-minutes guys like Collison on page 5: viewtopic.php?f=6&t=1314113&start=80

Wow, Doc MJ seems quite smart. Is this the typical general board discussion?

[Wannabe MEP]

Wow, Doc MJ seems quite smart. Is this the typical general board discussion?

General board? HELL no. But that's pretty typical for the advanced stats discussions, where Doc MJ is one of the mainstays.

[GMATCallahan]

A few thoughts on this…
1) 0 is not a bad score -- it's meant to be "average" for an NBA player. Slight negative isn't even bad, considering role and price. Amare at -5.84 and $21.7 million? THAT'S bad.
2) Low minute players are always hard to figure out, so I wouldn't take this to mean too much for those guys.
3) I don't know how they're doing priors on this. Engelmann has used slight negative values as the priors for rookies, so that may be the case here. I also don't know how many years of data they're using, although it appears to be weighted heavily for this season.

Plus-minus metrics are far from perfect, but one advantage is that they do tend to account for defense, something that traditional statistics and traditional basketball coverage often fail to account for (in terms of individual players). I imagine that if we possessed this kind of data for players from the past such as Charles Barkley and Chris Mullin, they'd prove to be quite overrated because defensively, they gave up so much of what they produced offensively.

And in that regard, Stoudemire goes without saying; at least Barkley constituted a great rebounder.